Category AI News

‘Subliminal Learning’ in AI Fine-Tuning Can Teach Bad Habits

Anthropic study reveals that common AI fine-tuning methods may introduce hidden biases and risks into models, impacting their performance. This 'subliminal learning' phenomenon could lead to unintended consequences in AI systems, highlighting the importance of ethical AI development practices.

Zuckerberg Predicts Disadvantage for Those Without AI Glasses

Mark Zuckerberg predicts that people without AI glasses will be at a disadvantage in the future as these glasses are seen as the ideal tool to merge physical and digital worlds. Zuckerberg emphasizes the importance of this technology for staying competitive.

Mark Zuckerberg Unveils Meta’s Plan for Self-Improving AI

Meta CEO Mark Zuckerberg announces plans for a research lab to develop AI models capable of learning with minimal human intervention, emphasizing a focus on self-improving and superintelligent AI. The goal is to create systems that can rapidly adapt and enhance their own capabilities, potentially leading to advancements in various fields.

GitHub Copilot Hits 20 Million Users Milestone

GitHub Copilot, a leading AI coding tool, gained an additional five million users in the past three months, bringing its total user base to over 20 million. This growth highlights the tool's popularity and significance in the coding community.

Manus to Introduce Major Upgrade for AI Agent Platform

Chinese startup Manus is set to launch a significant update to its AI agent platform, enabling collaborative research by assigning tasks to multiple agents simultaneously. This update is expected to be the biggest since the platform's inception, offering enhanced capabilities for researchers and developers. The feature will allow for more efficient and scalable AI projects, potentially impacting the landscape of AI research and development.

OpenAI Expands Stargate to Europe through Norwegian Deal

OpenAI partners with a new data center in Norway to establish its Stargate infrastructure in Europe, marking a significant expansion for the artificial intelligence company. This move positions OpenAI to better serve European customers and advance its AI capabilities in the region.

Lawsuit alleges Meta pirated porn to train AI while hiding piracy

A lawsuit claims Meta may have distributed pirated porn to minors under the guise of training AI algorithms. The lawsuit alleges Meta used the pirated content to train its AI systems while hiding its piracy activities. The lawsuit raises concerns about potential harm to minors and violations of copyright laws.

Google Chrome Introduces AI-Powered Store Summaries for US Shoppers

Google Chrome's new feature will provide shoppers with pop-up summaries detailing store reputation in areas like product quality, pricing, customer service, and returns. The AI-powered tool aims to assist consumers in making informed decisions while shopping online, enhancing their overall experience and confidence in their purchases.

Microsoft Edge Introduces AI ‘Copilot Mode’

Microsoft's Edge browser launches an AI-powered 'Copilot Mode' for enhanced web browsing, utilizing artificial intelligence to improve user experience. This feature aims to provide smarter suggestions and assistance while browsing, showcasing Microsoft's commitment to integrating AI technologies into their products.

Anthropic Faces Backlash for Imposing Rate Limits on Claude Users

Anthropic received criticism from developers for implementing weekly rate limits on some users of Claude Code, attributing the move to users running the code constantly. The decision sparked a negative reaction on social media platforms, with users expressing discontent over the new restrictions.